Metagenome-assembly-and-binning

Current workflows in the symbiosis department

Step 1 -- I have a bunch of reads, now what?

Metagenome sequencing reads

Check your reads with FastQC
Check what organisms are there: PhyloFlash for 16S rRNA and COI
Quality trim your reads. We often use a quality of 2 since the reads get a higher quality after error correction (Bbduk). Remove PhiX, TruSeq adaptors.
If your paired-end reads overlap, try merging the reads (Bbmerge).
HGV tip: Check the K-mer spectrum.
Assemble your reads (e.g. Megahit, MetaSPades, IDBA_UD)
- Megahit: Fast, but produces highly fragmented assemblies
- MetaSpades: Crashes often with more than one library. Slow
- IDBA_UD: Produces good N50, and has a good running time. The output does not include an assembly graph.

Step 2 -- Assembly

Before proceeding think about the following:

Does it make sense to pool the reads of different libraries together?

If yes, do you have access to a computer that can deal with all the data?

If you only have the sequencing of one sample, then assemble the one library. If your assembly is completely crappy, consider sequencing more samples.
Kmer 127: for reads >150 bp and coverage >5X
Brandon's script combines scaffold with graph

Step 3 -- Binning

Differential coverage and GC% (e.g. GBTools)
Taxonomy (Blobology, Phyla-Amphora)
Check RNA (e.g. Barrnap)
Linkage analysis (e.g. Bandage or Albertsen protocol)

Check quality control of the assembled genomes (See step 5)

Step 4 -- Reassembly

Often improves assembly metrics, but you should try it yourself.
More work and not scalable to many data sets.

Step 5 -- Quality control of assembly

CheckM (Prokaryotes)
BUSCO (Eukaryotes)
Quast

Optional:

Phylogenomic placement
ANI (e.g. Jspecies, or Kostas website)

Other resources

Stephen Turner's blog about genetic tips and other informatic news: http://www.gettinggeneticsdone.com/
Adrien's blog on his rambling about bioinformatic, include tutorials and scripts examples: http://aassie.net/Pro/
Nature computational biology journal collection of fundamental paper exaplaining computational concept (eg. Bayesian statistic, Debruijn graph, etc...): http://www.nature.com/nbt/collections/compbio/index.html

About this repository

This repository contains results from metagenome binning roundtable discussion moderated by Liz S on 26 Sep 2016. Sharing it as a Git repository has three aims - (i) make it possible to edit it collaboratively, (ii) for newbies, have some practice using Git and Github for a simple project, and (iii) take advantage of the Markdown formatting to start building a simple webpage that can be used as a reference on metagenome binning.

Formatted in Markdown - learn more about the syntax here.

Learn more about Git and Github from the Github help pages, and Symbiosis wiki (internal link)

Suggested workflow:

Get a Github account if you don't have one already.
Fork this repository and clone it to your local machine.
Edit the text, create new pages, insert links, images, etc.
Commit those changes. (Tip: Commit small "chunks" of changes at a time, so that they can be easily undone if necessary, and document your commits with short but descriptive commit messages.)
Push changes back to Github.
Submit a pull request for Liz to merge your changes with the master version of the repository.